## wpiea2026027-source-pdf

## Source details

**Canonical URL:** [wpiea2026027-source-pdf](https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026027-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2026/english/wpiea2026027-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2026/english/wpiea2026027-source-pdf.pdf.json)

---

### Demographic context and headline trends
- Old-age dependency ratio—estimated at 21.2 percent in 2024—is projected to double by 2041 (about 42 percent).
- Annual births declined to 9.5 million in 2024 from roughly 18 million in 2016.
- Working-age population (ages 20–60) expected to contract from 840 million to 614 million by 2050, corresponding to an average annual decline of about 1.2 percent.
- Urbanization increased from 17.2 to 64.6 percent over the past five decades.
- Population aging projection (model definition): old-age dependency ratio projected to increase from 21.2 percent in 2024 to about 42 percent in 2041; model-specific projection: old-age dependency ratio rises from 25.4 to 45.2 (population age 60+ over population age 20–59).

### Pension system structure, coverage, and parameters
- Three-pillar architecture:
  - Pillar 1 (public): Residents Pension Scheme (RPS) and Urban Employees’ Pension Scheme (UEPS).
  - Pillar 2: enterprise and occupational annuities (voluntary since 2004; mandatory occupational annuities for public-sector employees since 2014).
  - Pillar 3: voluntary private savings through tax-exempt accounts (up to 12,000 RMB per year), introduced nationally in 2022.
- Coverage and enrollment:
  - Near-universal old-age coverage extending pensions to over one billion people.
  - By 2023 approximately 1.07 billion citizens enrolled in public pension schemes: 521 million in UEPS and 545 million in RPS.
  - In 2023 only about 44 percent of China’s working-age population had effective coverage in an earnings-related scheme.
  - By 2022, about 71 million participants—roughly 9 percent of the workforce—enrolled in enterprise or occupational annuity schemes.
  - About 70 million Pillar 3 accounts opened since rollout.
  - Working population close to 900 million people in 2024 (aged 15-59 years).
- Spending, subsidies, and financing:
  - China spends 5.4 percent of GDP on public pensions (2023).
  - Government subsidies totaled 1.4 percent of GDP in 2023, with 78 percent directed to the urban scheme.
  - Approximately 26 percent of public pension expenditures are financed via government subsidies (reported with comparability caveats).
- Key actuarial and benefit parameters:
  - UEPS DB component: 1 percent of the average of the individual’s wage and the province-wide average earnings for each year of contribution (subject to minimum 15 years of contributions prior to 2024 reform).
  - Employees pay 8 percent of wages to notional individual accounts; contribution rate (employee + employer) 24 percent (8 percent by workers and 16 percent by employers).
  - UEPS average-wage worker replacement rate: 68.3 percent; half-average-wage worker: 87.3 percent; high earners: 58.8 percent.
  - RPS: minimum payment of RMB 143 per month; combined RPS benefits average about 3 percent of per-capita GDP; when no contributions made, RPS benefits amount to just 2 percent of the average wage.
  - Notional defined contribution (NDC) annuity divisor was last updated in 2005 and set at 139 months for retirement at age 60.
  - Life expectancy at age 60 has risen to 21.7 years (260.5 months) and is projected to increase to 25.5 years (305.8 months) by 2050.
- Historical reforms timeline (selected):
  - Dismantling of the "iron rice bowl" (late 1970s); 1997 multi-tier scheme; 2009–2014 coverage extension; 2014 RPS largely subsidized; 2015 integration of government employees (contributions: 8 percent employees, 16 percent employers); 2018 central adjustment fund pooling at least 3 percent of each province’s revenues; 2022 national-level pooling with cross-province transfers (~RMB 244 billion); 2024 retirement age reform and longer minimum contribution period.

### Model framework, calibration, and simulation design
- Model type: extended 80-cohort overlapping generations (OLG) framework (based on Cao et al. (2024)), closed-economy assumption, no endogenous human capital accumulation or migration.
- Population sectors: explicit urban–rural differentiation—urban households contribute to and receive earnings-related pensions; rural households receive fixed transfers without contributing.
- Simulation horizon: 240-year horizon starting from 2024, with focus on transitional dynamics through 2050.
- Key calibration values and fiscal assumptions:
  - Discount factor β = 0.995.
  - Government consumption ≈ 20 percent of GDP.
  - Labor preference γ = 2.4.
  - TFP growth γₐ = 0.04 (2024 potential growth 4.2).
  - Capital share α = 0.5; depreciation rate δ = 0.1.
  - Effective payroll tax τˢˢ = 0.083.
  - Urban pension expense over GDP (%) = 5; rural pension (urban-rural resident pension exp. over GDP (%)) = 0.4.
  - Replacement rate ξ = 0.44 calibrated to match urban worker pension spending of 5 percent of GDP.
  - Government debt calibrated at 110 percent of GDP (augmented definition, netting out social security funds).
  - Government debt path assumed fixed across scenarios: government debt increasing from 110 percent of GDP in 2024 to 150 percent within a decade before stabilizing.
  - Consumption tax rate adjusts endogenously period-by-period to maintain intertemporal budget balance.
- Model limitations noted:
  - Annual-frequency model implements phased statutory retirement age changes in discrete steps, producing observable jumps.
  - Does not incorporate flexible retirement behavior (may overstate macro-fiscal impacts).
  - Does not distinguish between genders in some simulations (may understate impacts for women groups).

### Simulation scenarios and headline macro-fiscal impacts (selected outcomes through 2050)
- Population aging (UN baseline, demographic-only):
  - Real GDP growth declines by approximately 2 percentage points between 2024 and 2050 due to shrinking labor force and lower capital accumulation.
  - GDP drops by 18 percent by 2050 (relative to baseline without reforms) in model-specific box estimates.
  - Pension expenditures surge from 5.4 percent to 15.3 percent of GDP by 2050.
  - Social security contributions as a ratio of GDP remain broadly flat in the baseline demographic-only scenario.
  - National household saving rate projected to decline from 28 percent in 2024 to 14 percent in 2050.
  - Interest rates fall by one percentage point due to lower growth prospects and reduced investment demand.
- Legislated 2024 retirement age reform (gradual increase by three years, phased 2025–2040):
  - Boosts GDP per capita by 5.6 percent by 2050; equivalent to lifting average annual growth by roughly 0.2 percentage points.
  - Annual real GDP growth rises modestly by approximately 0.2 percentage points (Δ GDP growth between 2024 and 2050: -2.0 | -1.8 | 0.2).
  - Pension expenditures decline from 15.3 percent to 11.9 percent of GDP by 2050 (reduction of 3.4 percent of GDP).
  - National saving rate increases by 0.9 percentage point in 2050 relative to the population aging scenario.
  - Interest rate increases marginally by 14 basis points.
- Additional reform scenarios (assessed relative to legislated reform):
  - Doubling RPS benefits:
    - Pension expenditures increase by 0.6 percent of GDP by 2050 compared to legislated baseline.
    - Rural elderly welfare: consumption-equivalent welfare gains exceeding 10 percent for some cohorts.
    - Saving effects: overall saving rate falls by 0.3 percentage points; rural household saving rate declines by 3.4 percentage points.
    - Consumption inequality falls by 0.3 percentage points.
  - Aligning UEPS NDC divisor with rising longevity (annuity divisor from 139 to 305 months; illustrative DB/NDC splits):
    - Baseline (NDC = one-third of total benefits): replacement rate reduced from 43 percent to 35.2 percent (an 8-percentage point decline); by 2050 GDP increases by 1.9 percent; pension expenditures decline by 1.3 percent of GDP; national saving rate increases by 0.7 percentage points.
    - Lower-bound (NDC = 3 percent of overall replacement rate): overall replacement rate falls to 41.4 percent (a 1.6 percentage point decline); by 2050 GDP increases by about 0.9 percent; pension expenditures decline by 0.7 percent of GDP; national saving rate rises by 0.3 percentage points.
  - Accelerating retirement age increase to 65 for all workers (by 2040):
    - By 2050, real GDP projected to be 3.1 percent higher relative to legislated policy.
    - Average annual growth boost of 0.1 percentage points.
    - Fiscal expenditures decline by an additional 1.8 percent of GDP.
    - Interest rate increases by 4 basis points.
  - Faster urbanization via increased rural youth migration (urbanization from 66 percent to 74 percent by 2050):
    - Lifts GDP by 6 percent (cumulative GDP per capita gain ~6 percent; average annual gain ~0.2 percent).
    - Pension expenditures reduced by 0.7 percent of GDP compared to legislated policy trajectory.
    - Caveat: newly urbanized workers retiring after decades could generate additional long-run fiscal pressures due to higher UEPS benefits.

### Quantitative indicators and tabulated scenario snapshots (selected exact values)
- Pension spending (percent of GDP, 2050) under scenarios (values as presented in the source panel):
  - 0.6 (+12.5) | -0.7 (+11.2) | -0.7 (+11.2) | -1.3 (+10.6) | -1.8 (+10.1)
- CAGR GDP (percent/yr, 2024-2050) across scenarios (values as presented in the source panel):
  - -0.01 | 0.20 | 0.03 | 0.06 | 0.11
- Model box projections (selected):
  - Urbanization rate: 66 percent in 2024 → 68 percent by 2050 (baseline); alternate scenario: to 74 percent by 2050.
  - Working-age population (ages 20–59) declines by 26.2 percent by 2050 (model box).
  - Old-age dependency: model box reports rise from 25.4 to 45.2 (population 60+/20–59).
  - National saving rate: projected to decline from 28 percent in 2024 to 14 percent in 2050 (baseline aging).
  - Interest rates: fall by one percentage point in baseline aging; rise by 14 basis points under legislated reform.

### Policy implications, trade-offs, and recommended levers
- Gradual retirement age increases:
  - Can materially mitigate fiscal pressures and support higher GDP through greater labor supply and higher national saving.
  - Legislated 2024 reform reduces pension expenditures and raises GDP per capita but does not eliminate long-run spending pressures.
- Benefit rebalancing:
  - Doubling RPS benefits improves welfare for rural elderly (some cohorts >10 percent consumption-equivalent gains) but raises pension expenditures and lowers saving rates—trade-offs between equity and fiscal sustainability.
- Indexing pension parameters to longevity:
  - Updating the NDC annuity divisor to reflect rising life expectancy reduces replacement rates, lowers pension expenditures, and raises GDP and national saving.
  - Impact magnitude depends on DB vs NDC share; results presented as ranges.
- Encouraging labor reallocation and urbanization:
  - Faster urbanization supports GDP growth and reduces pension expenditures in the medium term; long-run fiscal pressures may rise as urbanized cohorts access more generous UEPS benefits.
- Complementary reforms emphasized:
  - Strengthening rural pension benefits; broadening participation in urban scheme via hukou reform; adjusting pension parameters for rising life expectancy; accelerating retirement age increases; central pooling and inter-provincial transfers (central fund pooling at least 3 percent of provincial revenues; national pooling and transfers ~RMB 244 billion in 2022).

### Box highlight — 2024 Retirement Age Reform and media sentiment (selected facts)
- Legislative changes implemented 2025–2040:
  - Male employees: from age 60 to age 63 (by 3 months per year).
  - Female blue-collar employees: from age 50 to age 55 (by 6 months per year).
  - Female white-collar employees: from age 55 to age 58 (by 3 months per year).
  - Minimum contribution period increases from 15 to 20 years between January 1, 2030 and 2039 (increment of six months per year).
  - Early retirement: employees meeting minimum contribution period may opt for flexible early retirement up to three years without actuarial deductions.
- Media and stakeholder sentiment (September 2024 legislative month):
  - Household average sentiment: -2 (less negative than international comparisons average: -3).
  - Government and researchers exhibited relatively positive sentiment.
  - Media analysis based on Factiva database covering official state media and other widely read economic and financial outlets.
- Calibration and box-specific parameters:
  - Discount factor β = 0.995; labor preference γ = 2.4; TFP growth γₐ = 0.04; capital share α = 0.5; depreciation δ = 0.1; effective payroll tax τˢˢ = 0.083.
  - Government debt path: 110 percent of GDP in 2024 → 150 percent within a decade before stabilizing.
  - Urban pension expense over GDP (%) = 5; rural pension = 0.4 percent of GDP.
  - Replacement rate ξ = 0.44 calibrated to match urban worker pension spending of 5 percent of GDP.

*Source: IMF Working Paper — "Population Aging and Pension Reforms in China", Working Paper No. WP/2026/027 (chapter excerpts and Box 1).*

### 1. Population Structure ................................................................................................

### 1. Population Structure

### Demographic context and trends
- Old-age dependency ratio—estimated at 21.2 percent in 2024—is projected to double by 2041.  
- Annual births declined to 9.5 million in 2024 from roughly 18 million in 2016.  
- Working-age population (ages 20–60) expected to contract from 840 million to 614 million by 2050, corresponding to an average annual decline of about 1.2 percent.  
- Urbanization increased from 17.2 to 64.6 percent over the past five decades.

### Pension system evolution and structural features
- China’s pension architecture now consists of three pillars: a broad-based public system (Pillar 1), limited occupational pensions (Pillar 2), and emerging private savings schemes (Pillar 3).  
- Reforms highlighted: dismantling of the "iron rice bowl" (late 1970s), Residents Pension Scheme (RPS) and Urban Employees’ Pension Scheme (UEPS) establishment, 2015 integration of government employees into the urban system, reforms in 2018 and 2022, and the 2024 retirement age reform.  
- Coverage: near-universal old-age coverage achieved, extending pensions to over one billion people.  
- Remaining gaps: relatively low retirement ages (particularly for women) by OECD/regional standards; uneven benefit adequacy (urban schemes relatively generous; rural pensions modest); some pension parameters not linked to rising life expectancy; high minimum contribution requirements constraining urban-scheme participation; persistent fragmentation and portability barriers.

### Model framework and calibration
- Dynamic model: extended 80-cohort overlapping generations (OLG) framework based on Cao et al. (2024).  
- Economy: closed economy assumption; no endogenous human capital accumulation or migration.  
- Population sectors: explicit urban–rural differentiation—urban households contribute to and receive earnings-related pensions; rural households receive fixed transfers without contributing.  
- Simulation horizon: 240-year horizon starting from 2024, with focus on transitional dynamics through 2050.

### Simulation scenarios and headline macro-fiscal impacts
- Population aging (UN population projections baseline):  
  - Between 2024 and 2050, real GDP growth declines by approximately 2 percentage points in the model due to the shrinking labor force and resulting decline in capital accumulation.  
  - Pension expenditures surge from 5.4 percent to 15.3 percent of GDP by 2050, driven by a rising old-age dependency ratio.  

- Legislated 2024 retirement age reform (gradual increase by three years):  
  - Boosts GDP by 5.6 percent by 2050, equivalent to lifting average annual growth by roughly 0.2 percentage points.  
  - Reduces pension expenditures from 15.3 percent to 11.9 percent of GDP by 2050.  
  - National saving increases as urban workers delay pension access and raise savings; the size of the retired workforce that dissaves is smaller.

- Additional reform scenarios (compared with legislated policy):  
  - Doubling RPS benefits: raises pension expenditure marginally by 0.6 percent of GDP by 2050; significantly improves consumption-equivalent welfare for elderly rural residents—in some cohorts by more than 10 percent—and narrows overall consumption inequality; the saving rate falls, especially among rural households.  
  - Aligning UEPS Notional Defined Contribution (NDC) divisor with rising longevity: lowers the replacement rate for urban pensions by 8 percentage points; raises GDP by 2 percent; reduces pension expenditures by up to 1.3 percent of GDP.  
  - Accelerating retirement age increase to 65 for all workers: boosts GDP by 3 percent; reduces pension spending by an additional 1.8 percent of GDP by 2050.  
  - Faster urbanization via increased rural youth migration (raising urbanization from 66 to 74 percent): lifts GDP by 6 percent.

### Key policy implications emphasized
- Gradual retirement age increases can materially mitigate fiscal pressures and support higher GDP through greater labor supply and higher national saving.  
- Benefit rebalancing—e.g., increasing RPS benefits—can improve welfare for rural elderly but has fiscal and saving-rate implications.  
- Indexing pension formulas to longevity (NDC divisor alignment) can reduce replacement rates, lower pension expenditures, and raise GDP.  
- Encouraging labor reallocation toward urban areas can support GDP growth and interact with pension sustainability outcomes.

*Source: IMF Working Paper — Population Aging and Pension Reforms in China (chapter: "1. Population Structure").*

### 0.7 percent of GDP by 2050.5F

### 0.7 percent of GDP by 2050.5F

### Structural transformation and long-run fiscal pressures
- Rapid urbanization and demographic change may introduce long-run fiscal pressures as newly urbanized workers eventually reach retirement.
- Urbanization rate of 74 percent is consistent with projections by Liu et al. (2017) and Pathak Raimedhi (2023).
- Population aging projection: old-age dependency ratio projected to increase from 21.2 percent in 2024 to about 42 percent in 2041.

### Literature, methods, and contribution
- Overlapping generations (OLG) models are central to assessing macroeconomic consequences of demographic change; Zhai (2024) estimates population aging could reduce China’s average annual growth by 1.1–1.4 percentage points over five decades.
- This paper adapts a state-of-the-art OLG model to explicitly account for China-specific urban-rural heterogeneity in demographics, labor markets, and pension structures.
- The paper compares China’s pension system to international peers and quantifies macro-fiscal impacts of reform options including retirement age increases, benefit adjustments, and urbanization strategies.

### Evolution of China’s public pension system (1949–2024)
- 1949–1978: "iron rice bowl" SOE-based pensions; many rural workers uncovered.
- 1997: multi-tier pension scheme launched for urban salaried private-sector employees.
- 2009–2014: pension coverage extended to rural residents and non-salaried urban workers; 2014 Residents Pension Scheme (RPS) largely subsidized by government, near-universal coverage but low benefits for rural residents and non-salaried urban workers.
- 2015: integration of government employees into urban system; government employees required to contribute 8 percent of salaries, employers 16 percent (funds for private and public sector workers remain separate).
- 2018: central adjustment fund for urban employees’ pension scheme implemented, pooling at least 3 percent of each province’s revenues.
- 2022: evolved into a full national-level pooling system with substantial cross-province transfers (~RMB 244 billion), a unified pension data and risk-control platform, and nationwide operating standards.
- 2024 reform: legislated gradual retirement age increase and longer minimum contribution period:
  - Retirement age of the urban employees’ pension scheme increased by three to five years depending on gender and job type, phased in between 2025 to 2040.
  - Minimum contribution period increased from 15 to 20 years.
  - Individuals not fulfilling the criterion may pay a lump sum (procedures and transitional rules apply for those who first participated before July 1, 2011).

### Three-pillar pension architecture and key parameters
- Pillar 1 (public):
  - Composed of Residents Pension Scheme (RPS) and Urban Employees’ Pension Scheme (UEPS).
  - RPS: voluntary for those not covered by UEPS, heavily subsidized, near-universal participation; non-contributory DB component pays a flat pension differing across provinces, with a minimum payment of RMB 143 per month.
    - Combined RPS benefits average about 3 percent of per-capita GDP.
    - For some retirees who have not contributed throughout their lifetime, outcomes resemble a non-contributory pension scheme (pillar 0).
  - UEPS: mandatory for urban employees (public and private); DB component pays 1 percent of the average of the individual’s wage and the province-wide average earnings for each year of contribution (subject to minimum 15 years of contributions prior to 2024 reform).
    - Employees pay 8 percent of wages to notional individual accounts.
    - Average UEPS benefit (DB + NDC) remains high at over 40 percent of per-capita GDP (more than 10 times the RPS benefit).
  - Notional defined contribution (NDC) annuity divisor was last updated in 2005 and set at 139 months for retirement at age 60.
    - Life expectancy at age 60 has risen to 21.7 years (260.5 months) and is projected to increase to 25.5 years (305.8 months) by 2050.
- Pillar 2:
  - Voluntary enterprise annuities (defined contribution) since 2004; mandatory occupational annuities for public-sector employees since 2014.
  - By 2022, about 71 million participants—roughly 9 percent of the workforce—enrolled in enterprise or occupational annuity schemes.
- Pillar 3:
  - Introduced nationally in 2022; voluntary individual private savings through tax-exempt accounts (up to 12,000 RMB per year).
  - About 70 million accounts opened since rollout.
  - Working population close to 900 million people in 2024 (aged 15-59 years).

### Benefit adequacy, progressivity, and contribution rates
- UEPS theoretical replacement rates and contribution structure:
  - Average-wage worker replacement rate: 68.3 percent.
  - Half-average-wage worker replacement rate: 87.3 percent.
  - High earners replacement rate: 58.8 percent.
  - Contribution rate: 24 percent (8 percent by workers and 16 percent by employers).
  - OECD average contribution rate: 18.2 percent.
  - UEPS is more progressive than OECD peers due to DB component weighting regional average wage.
- RPS adequacy:
  - When no contributions are made, RPS benefits amount to just 2 percent of the average wage—among the lowest across international benchmarks.
  - Regional comparators (Thailand, Sri Lanka, Vietnam) offer targeted non-contributory pensions ranging from 3 to 8 percent of the average wage; advanced economies offer more than 15 percent of the average wage.
- Coverage and spending:
  - By 2023 approximately 1.07 billion citizens enrolled in public pension schemes: 521 million in UEPS and 545 million in RPS.
  - In 2023 only about 44 percent of China’s working-age population had effective coverage in an earnings-related scheme (compared to over 70 percent in Japan and Korea).
  - China currently spends 5.4 percent of GDP on public pensions.
  - Government subsidies totaled 1.4 percent of GDP in 2023, with 78 percent directed to the urban scheme.
  - Effectively about 26 percent of public pension expenditures are financed via government subsidies (noting comparability caveats with European studies).

### International comparison and policy implications underscored in the text
- Retirement age:
  - China’s retirement ages remain low relative to OECD peers, especially for blue-collar women.
  - Over the next 15 years, retirement age for women is set to rise to 55 for blue-collar occupations and to 58 for white-collar positions; men’s retirement age rising from 60 to 63.
  - International examples: United States legislated a gradual increase to full retirement age 67 in 1983 when its old-age dependency ratio was around 15 percent; Australia announced increase to age 67 in 2009 at an early demographic stage; Korea transitioned from 62 to 65 as old-age dependency ratio crossed 20 percent.
- NDC annuity divisor and longevity:
  - Outdated annuity divisor (139 months set in 2005) understates current life expectancy, effectively rendering the system more generous as longevity rises.
  - International good practice links the annuity divisor to life expectancy at retirement with regular updates (examples: Sweden, Italy, Poland).
- Fiscal and distributional considerations:
  - China’s pension spending at 5.4 percent of GDP is broadly comparable to countries with similar old-age dependency ratios (e.g., United States, Canada, Australia allocating between 4.0 and 6.7 percent of GDP), but China’s faster aging implies mounting pension demands over a shorter time horizon.
  - Government subsidies (1.4 percent of GDP in 2023) may have regressive impacts since main beneficiaries are often middle- and high-income groups.

### Quantitative and policy focus areas highlighted
- Model and simulation focus:
  - The paper presents simulations under alternative assumptions for the share of total pensions attributable to DB and NDC components, given uncertainty about the exact DB/NDC split.
  - Updating the NDC divisor has impacts that depend on assumptions about the DB vs. NDC share—results are presented as a range in the simulation section.
- Key reform levers discussed:
  - Gradual increases in retirement ages (as legislated in 2024).
  - Revisions to benefit parameters including updating the NDC annuity divisor to reflect rising life expectancy.
  - Measures to increase earnings-related coverage and strengthen Pillar 2 and Pillar 3 uptake.
  - Central pooling and inter-provincial transfers to address provincial disparities (central fund pooling at least 3 percent of provincial revenues; national pooling and transfers ~RMB 244 billion in 2022).

*Source: IMF Working Paper excerpt (Population Aging and Pension Reforms in China).*

### Box 1  2024 Retirement Age Reform and Media Sentiment

### Box 1  2024 Retirement Age Reform and Media Sentiment

### Legislative changes in the 2024 reform
- Statutory retirement age increases begin in 2025 and phase in over a 15-year period:
  - Male employees: from age 60 to age 63 (by 3 months per year).
  - Female blue-collar employees: from age 50 to age 55 (by 6 months per year).
  - Female white-collar employees: from age 55 to age 58 (by 3 months per year).
- Minimum contribution period for a monthly basic pension:
  - Starting January 1, 2030, progressively increases from 15 to 20 years until 2039 (increment of six months per year).
  - Employees who reach statutory retirement age without meeting the required contribution period may extend contributions or make a one-time payment to fulfill the requirement.
- Early retirement provisions and floor on retirement ages:
  - Employees who have met the minimum contribution period may opt for flexible early retirement, with a maximum early retirement period of three years (without actuarial deductions).
  - Retirement age cannot fall below original statutory limits: 50 or 55 years for female employees and 60 years for male employees.

### Media coverage and stakeholder sentiment
- Media activity:
  - The 2024 reform sparked intensified media coverage in September 2024, with the frequency of mentions of pension reforms in the print media surpassing that of any other year since 2010 (see Figure B1).
  - Plans to raise the retirement age were also proposed by the government in 2012 and 2013, which generated heightened media debate but were not adopted then.
- Stakeholder positions in print media during the legislative month of September 2024:
  - Government and researchers exhibited relatively positive sentiments toward the reform.
  - Households adopted a more critical stance, with average sentiment of -2 (Figure B2).
  - The household sentiment in China (average -2) was notably less negative than that observed in international comparisons (average -3).
- Data and methodology note:
  - Sources: Staff estimates based on Factiva database.
  - Analysis is based on a wide sample of news articles from Factiva, covering official state media and other widely read economic and financial outlets, reflecting the diversity of China’s media environment within a comparatively strong regulatory framework.

### Context in the macroeconomic model and calibration assumptions
- Model overview:
  - The analysis employs an annual dynamic OLG model tailored to China with 80 overlapping generations, distinguishing urban and rural sectors and calibrated to China’s 2024 demographic profile.
- Key calibration and fiscal assumptions:
  - Discount factor β = 0.995 (consumption over GDP targeted at approximately 40 percent).
  - Government consumption calibrated at approximately 20 percent of GDP.
  - Labor preference γ = 2.4 (labor supply targeted at 0.33).
  - TFP growth γₐ = 0.04 (2024 potential growth 4.2).
  - Capital share α = 0.5; depreciation rate δ = 0.1.
  - Effective payroll tax τˢˢ = 0.083.
  - Urban pension expense over GDP (%) = 5; rural pension (urban-rural resident pension exp. over GDP (%)) = 0.4.
  - Replacement rate ξ = 0.44 calibrated to match urban worker pension spending of 5 percent of GDP.
  - Government debt calibrated at 110 percent of GDP (augmented definition, netting out social security funds).
  - Government debt path assumed fixed across scenarios: government debt increasing from 110 percent of GDP in 2024 to 150 percent within a decade before stabilizing.
  - To maintain intertemporal budget balance, the consumption tax rate adjusts endogenously on a period-by-period basis.

### Quantitative results and projections (selected simulation outcomes through 2050)
- Demographic and labor:
  - Urbanization rate increases from 66 percent in 2024 to 68 percent by 2050.
  - Working-age population (ages 20–59) declines by 26.2 percent by 2050.
  - Old-age dependency ratio rises from 25.4 to 45.2 (model definition: population age 60+ over population age 20–59).
- Output and growth:
  - GDP drops by 18 percent by 2050 (relative to baseline without reforms).
  - Annual real GDP growth decelerates by nearly 2 percentage points by 2050.
- Public finances and pensions:
  - Pension expenditures increase from 5.4 percent to 15.3 percent of GDP by 2050.
  - Social security contributions as a ratio of GDP remain broadly flat in the baseline demographic-only scenario.
- Savings and interest rates:
  - National household saving rate projected to decline from 28 percent in 2024 to 14 percent in 2050.
  - Interest rates fall by one percentage point due to lower growth prospects and reduced investment demand.
- Model horizon and interpretation:
  - Simulations run over a 240-year horizon; results focused on transition dynamics through 2050.
  - Results incorporate updated UN demographic projections and assume fixed NDC annuity factors in baseline estimates.
  - Pension projections assume generally constant pension coverage rates and contribution careers; lack of microdata implies simplified pension projections.

*Source: IMF staff estimates and model simulations as presented in Box 1, “2024 Retirement Age Reform and Media Sentiment,” IMF Working Paper.*

### 2. Assessment of the 2024 Retirement Age Reform

### 2. Assessment of the 2024 Retirement Age Reform

### Simulation design and caveats
- Simulation adds a phased increase in the statutory retirement age of three years for urban households over the next 15 years, concluding in 2040.
- Because the model has annual frequency, increases are implemented in discrete steps in years 5, 10, and 15, producing observable jumps in the policy variable along the simulation path.
- Model limitations noted:
  - The model does not incorporate flexible retirement behavior; some individuals may still retire earlier than the statutory age, so macro-fiscal impacts may be somewhat overstated.
  - The model does not distinguish between genders; for certain groups of women—whose statutory retirement age increases more under the reform—the model may understate the true impact of the reform.

### Macroeconomic and fiscal impacts of the legislated reform (comparison with population aging scenario)
- Key outcomes by 2050:
  - Annual real GDP growth rises modestly by approximately 0.2 percentage points.
  - GDP per capita is 5.6 percent higher by 2050.
  - Pension expenditures decline by an estimated 3.4 percent of GDP—from 15.3 percent to 11.9 percent of GDP—by 2050.
  - National saving rate increases by 0.9 percentage point in 2050, relative to the population aging scenario.
  - Interest rate increases marginally by 14 basis points.
- Tabulated indicators (Population Aging (A) vs Legislated Reform (B) and difference (B)-(A)):
  - Pension spending (percent of GDP): 15.3 | 11.9 | -3.4
  - Δ GDP growth between 2024 and 2050 (percent): -2.0 | -1.8 | 0.2

### Heterogeneous household effects and channels
- Two countervailing effects on savings:
  - Longer working lives reduce saving motives for younger workers (shorter retirement to finance).
  - Retirees have lower saving rates than workers; a smaller retiree share increases aggregate saving.
- Net effect in the legislated reform: composition effect dominates, raising the national saving rate by 0.9 percentage point by 2050.
- Labor supply expansion and higher investment demand tighten capital markets, pushing interest rates up (14 basis points).

### Additional reform scenarios (assessed relative to 2024 legislated reform)
- Overview of four complementary measures analyzed:
  - Enhancing benefits under the RPS (rural pension system).
  - Adjusting the UEPS NDC component to reflect rising life expectancy (reduce replacement rate).
  - Accelerating the increase of the retirement age (to 65).
  - Promoting faster urbanization (rural youth migration to cities).

- Panel summary indicators (Pension spending (percent of GDP, 2050) and CAGR GDP (percent/yr, 2024-2050)) across scenarios (presented as values for each scenario in order as shown in source):
  - Pension spending (percent of GDP, 2050): 0.6 (+12.5) | -0.7 (+11.2) | -0.7 (+11.2) | -1.3 (+10.6) | -1.8 (+10.1)
  - CAGR GDP (percent/yr, 2024-2050): -0.01 | 0.20 | 0.03 | 0.06 | 0.11

a) Improving rural resident pension benefits (doubling RPS benefits)
- Fiscal and macro effects:
  - Pension expenditures increase by 0.6 percent of GDP by 2050 compared to the legislated baseline.
  - GDP per capita falls modestly—by about 0.5 percent by 2050—due to a modest decline in labor supply relative to baseline.
- Distributional and welfare effects:
  - Rural households near retirement see consumption-equivalent welfare gains exceeding 10 percent.
  - Urban households experience minor welfare losses due to slightly higher consumption taxes needed to finance the expansion.
  - Overall consumption inequality falls by 0.3 percentage points.
- Savings and interest rate:
  - Both rural and urban household saving rates decline, with an overall drop of 0.3 percentage points.
  - Rural household saving rate declines by 3.4 percentage points.
  - Interest rate increases marginally by around 1 basis point in the long run.

b) Aligning the NDC (notional defined contribution) component with rising life expectancy (adjust annuity divisor from 139 to 305 months)
- Two illustrative simulations given uncertainty over DB/NDC split:
  - Baseline scenario (NDC = one-third of total pension benefits):
    - Overall replacement rate reduced from 43 percent to 35.2 percent (an 8-percentage point decline).
    - By 2050: GDP increases by 1.9 percent; pension expenditures decline by 1.3 percent of GDP; national saving rate increases by 0.7 percentage points; interest rate rises as capital-to-labor ratio adjusts upward.
  - Lower-bound scenario (NDC contributes 3 percent of overall replacement rate):
    - Adjusted overall replacement rate falls to 41.4 percent (a 1.6 percentage point decline).
    - By 2050: GDP increases by about 0.9 percent; pension expenditures decline by 0.7 percent of GDP; national saving rate rises by 0.3 percentage points.

c) Accelerating the increase of retirement age (statutory age to 65 by 2040)
- Design:
  - Statutory retirement age increases to 65 by 2040 with additional increments in years 7 and 13 of the simulation (two years beyond current reform trajectory).
  - Reform expanded to increase retirement age for RPS participants.
- Outcomes by 2050 relative to legislated policy:
  - Real GDP projected to be 3.1 percent higher.
  - Average annual growth boost of 0.1 percentage points.
  - Fiscal expenditures decline by an additional 1.8 percent of GDP.
  - Net national saving rate increases (composition effect dominates).
  - Interest rate increases by 4 basis points.

d) Faster urbanization through rural youth migration
- Implementation in model:
  - Assume half of the rural new workforce transitions to urban status by 2050 via birth-rate adjustments:
    - 휂_new,t_r = 0.5 휂_old_r
    - 휂_new_u = 0.5 휂_old_r + 휂_old_u
  - Urbanization rate rises from 66 percent to 74 percent.
- Impacts:
  - Structural transformation boosts cumulative GDP per capita by about 6 percent—equivalent to an average annual gain of 0.2 percent.
  - Pension expenditures reduced by 0.7 percent of GDP compared to the legislated policy trajectory.
  - Caveat: as newly urbanized individuals retire after 40 years, higher UEPS benefits could generate additional long-run fiscal pressures.

### Conclusion (from the chapter)
- China’s demographic transition presents significant challenges to long-term growth, fiscal sustainability, and social equity.
- The 2024 retirement age adjustment helps mitigate fiscal pressures but rapid increases in the old-age dependency ratio will drive significant pension spending increases without further reforms.
- Model simulations suggest timely, complementary reforms—(1) strengthening rural pension benefits; (2) broadening participation in urban scheme via hukou reform; (3) adjusting pension parameters for rising life expectancy; and (4) accelerating retirement age increases—can reduce fiscal pressures and help build a more resilient, inclusive pension system.

*Source: IMF Working Paper — "Population Aging and Pension Reforms in China", Working Paper No. WP/2026/027, Chapter 2.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026027-source-pdf.pdf_